Why PDFs get huge — and how to actually shrink them

A PDF containing nothing but text is tiny — a full novel fits in well under a megabyte. When a PDF is 20, 40 or 100 MB, the weight is almost always images: scanned pages, embedded photos, or graphics saved at print resolution. Understanding that is the whole trick to shrinking one.

Where the megabytes hide

• Scanned documents. A scanner turns each page into a photograph — often a 300–600 DPI image, which is 8–35 MB per page uncompressed. A 10-page scanned contract can easily outweigh a 300-page text PDF.
• Embedded photos. Reports and presentations exported to PDF carry every image at its original resolution, even when it’s displayed at a quarter of that size on the page.
• Fonts and duplicates. Every embedded font adds a little; badly-generated PDFs sometimes embed the same image or font repeatedly. Minor compared to images, but it adds up in long documents.

What compression actually does

A PDF compressor works on those embedded images: re-encoding them at a lower quality, downscaling ones that are larger than their display size, and re-rendering scan-pages at a sensible resolution. Text and vector content are usually left untouched — they’re already small and compress losslessly.

That leads to the honest version of what to expect:

• Scans and photo-heavy documents shrink dramatically — 50–90% reductions are normal.
• Text-only PDFs barely shrink at all. If a PDF is already mostly text, there’s nothing heavy to squeeze. A good tool tells you this instead of handing back a file 2% smaller (or worse, bigger).

Our PDF compressor recompresses embedded photos on normal pages while keeping the text selectable, and rasterizes pure scan-pages where that’s the big win. If compressing wouldn’t help, it keeps your original.

The trade-offs to know about

• Selectable text vs. scanned pages. On pages with a real text layer, compression that only touches images keeps text selectable and searchable. Pages that are already just pictures of text (scans) stay pictures — only cheaper ones.
• Form fields and bookmarks. Aggressive restructuring can flatten interactive forms and drop document outlines. If you need a fillable form to stay fillable, compress before filling it, or keep the original alongside.
• Print quality. Compression tuned for email (e.g. images at ~150 DPI) looks perfect on screen but slightly soft in print. For documents headed to a printer, choose a higher-quality setting.

Sometimes the better fix isn’t compression

• Only need a few pages? Split the PDF and send just those. Ten pages out of a 200-page scan beats any compressor.
• Building the PDF yourself from photos? Compress the images before they go in — images-to-PDF with per-image compression gives you a target size up front instead of a rescue job afterwards.
• Recurring reports? Fix the export settings at the source (image DPI/quality in Word, PowerPoint or your design tool) and the PDFs come out small every time.

A note on uploading contracts to “free PDF compressors”

The documents people most need to shrink — contracts, statements, medical records, ID scans — are exactly the ones that shouldn’t be uploaded to an unknown server. Compression doesn’t require a server: a browser can rebuild a PDF locally. Everything linked in this guide runs entirely on your device.